From: Bart Van Assche <bart.vanassche-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
To: Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
Cc: Christoph Hellwig <hch-jcswGhMUV9g@public.gmane.org>,
Sagi Grimberg <sagi-NQWnxTmZq1alnMjI0IkVqw@public.gmane.org>,
Laurence Oberman
<loberman-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>,
"linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org"
<linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org>
Subject: [PATCH 11/11] IB/srp: Prevent mapping failures
Date: Fri, 22 Apr 2016 14:16:31 -0700 [thread overview]
Message-ID: <571A94AF.7000609@sandisk.com> (raw)
In-Reply-To: <571A936F.7040409-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
If both max_sectors and the queue_depth are high enough it can
happen that the MR pool is depleted temporarily. This causes
the SRP initiator to report mapping failures. Although the SRP
initiator recovers from such mapping failures, prevent that
this can happen by limiting max_sectors.
Reported-by: Laurence Oberman <loberman-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
Signed-off-by: Bart Van Assche <bart.vanassche-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
Cc: Christoph Hellwig <hch-jcswGhMUV9g@public.gmane.org>
Cc: Sagi Grimberg <sagi-NQWnxTmZq1alnMjI0IkVqw@public.gmane.org>
---
drivers/infiniband/ulp/srp/ib_srp.c | 20 ++++++++++++++++++++
1 file changed, 20 insertions(+)
diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c
index a173ec4..ebd4d90 100644
--- a/drivers/infiniband/ulp/srp/ib_srp.c
+++ b/drivers/infiniband/ulp/srp/ib_srp.c
@@ -3172,6 +3172,7 @@ static ssize_t srp_create_target(struct device *dev,
struct srp_device *srp_dev = host->srp_dev;
struct ib_device *ibdev = srp_dev->dev;
int ret, node_idx, node, cpu, i;
+ unsigned int max_max_sectors;
bool multich = false;
target_host = scsi_host_alloc(&srp_template,
@@ -3228,6 +3229,25 @@ static ssize_t srp_create_target(struct device *dev,
target->sg_tablesize = target->cmd_sg_cnt;
}
+ if (srp_dev->use_fast_reg || srp_dev->use_fmr) {
+ /*
+ * FR and FMR can only map one HCA page per entry. If the
+ * start address is not aligned on a HCA page boundary two
+ * entries will be used for the head and the tail although
+ * these two entries combined contain at most one HCA page of
+ * data. Hence the "- 1" in the calculation below.
+ */
+ max_max_sectors = (srp_dev->max_pages_per_mr - 1) <<
+ (ilog2(srp_dev->mr_page_size) - 9);
+ if (target->scsi_host->max_sectors > max_max_sectors) {
+ shost_printk(KERN_WARNING, target->scsi_host,
+ PFX "Reducing max_sectors from %d to %d\n",
+ target->scsi_host->max_sectors,
+ max_max_sectors);
+ target->scsi_host->max_sectors = max_max_sectors;
+ }
+ }
+
target_host->sg_tablesize = target->sg_tablesize;
target->mr_pool_size = target->scsi_host->can_queue;
target->indirect_size = target->sg_tablesize *
--
2.8.1
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
next prev parent reply other threads:[~2016-04-22 21:16 UTC|newest]
Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-04-22 21:11 [PATCH 0/11] More SRP patches for kernel v4.7 Bart Van Assche
[not found] ` <571A936F.7040409-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-04-22 21:12 ` [PATCH 01/11] IB/srp: Fix a spelling error in a source code comment Bart Van Assche
[not found] ` <571A93AA.9000605-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-03 9:26 ` Christoph Hellwig
2016-04-22 21:12 ` [PATCH 02/11] IB/srp: Fix a comment Bart Van Assche
[not found] ` <571A93CF.9050506-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-04-26 20:52 ` Sagi Grimberg
2016-05-03 9:26 ` Christoph Hellwig
2016-04-22 21:13 ` [PATCH 03/11] IB/srp: Document srp_map_data() return value Bart Van Assche
[not found] ` <571A93E5.7030807-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-03 9:27 ` Christoph Hellwig
2016-04-22 21:13 ` [PATCH 04/11] IB/srp: Fix srp_map_data() error paths Bart Van Assche
[not found] ` <571A93FF.8020102-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-04-26 20:53 ` Sagi Grimberg
2016-05-03 9:28 ` Christoph Hellwig
2016-04-22 21:13 ` [PATCH 05/11] IB/srp: Introduce target->mr_pool_size Bart Van Assche
[not found] ` <571A9415.4090002-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-03 9:28 ` Christoph Hellwig
2016-04-22 21:14 ` [PATCH 06/11] IB/srp: Avoid that mapping failure triggers an infinite loop Bart Van Assche
[not found] ` <571A9427.8010306-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-03 9:29 ` Christoph Hellwig
2016-04-22 21:14 ` [PATCH 07/11] IB/srp: Move code out of a loop Bart Van Assche
[not found] ` <571A9443.6000600-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-04-26 21:00 ` Sagi Grimberg
2016-05-03 9:29 ` Christoph Hellwig
2016-04-22 21:15 ` [PATCH 08/11] IB/srp: Move common code into the caller Bart Van Assche
[not found] ` <571A9458.3050904-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-04-26 21:00 ` Sagi Grimberg
2016-05-03 9:29 ` Christoph Hellwig
2016-04-22 21:15 ` [PATCH 09/11] IB/srp: Fix a NULL pointer dereference Bart Van Assche
[not found] ` <571A9472.5050202-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-04-26 21:04 ` Sagi Grimberg
[not found] ` <571FD7F4.4090006-NQWnxTmZq1alnMjI0IkVqw@public.gmane.org>
2016-04-27 6:20 ` Leon Romanovsky
[not found] ` <20160427062053.GK7974-2ukJVAZIZ/Y@public.gmane.org>
2016-04-27 14:22 ` Bart Van Assche
2016-04-27 15:39 ` Bart Van Assche
2016-04-27 19:50 ` Bart Van Assche
2016-05-03 9:30 ` Christoph Hellwig
[not found] ` <20160503093032.GJ19931-jcswGhMUV9g@public.gmane.org>
2016-05-03 20:57 ` Bart Van Assche
[not found] ` <5729109E.6040500-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-03 21:01 ` Laurence Oberman
2016-04-22 21:16 ` [PATCH 10/11] IB/srp: Do not register memory if register_always = -1 Bart Van Assche
[not found] ` <571A9490.9010003-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-03 9:30 ` Christoph Hellwig
2016-04-22 21:16 ` Bart Van Assche [this message]
[not found] ` <571A94AF.7000609-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-04-23 22:12 ` [PATCH 11/11] IB/srp: Prevent mapping failures Laurence Oberman
2016-04-24 8:35 ` Leon Romanovsky
[not found] ` <20160424083538.GF7974-2ukJVAZIZ/Y@public.gmane.org>
2016-04-24 22:44 ` Laurence Oberman
2016-04-25 0:48 ` Bart Van Assche
[not found] ` <571D6975.2020905-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-04-25 5:54 ` Leon Romanovsky
[not found] ` <20160425055433.GG7974-2ukJVAZIZ/Y@public.gmane.org>
2016-04-25 9:46 ` Sagi Grimberg
[not found] ` <571DE793.6090705-NQWnxTmZq1alnMjI0IkVqw@public.gmane.org>
2016-04-25 15:53 ` Leon Romanovsky
[not found] ` <20160425155320.GH7974-2ukJVAZIZ/Y@public.gmane.org>
2016-04-25 16:16 ` Bart Van Assche
[not found] ` <571E42D2.70705-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-04-26 14:12 ` Leon Romanovsky
[not found] ` <20160426141229.GI7974-2ukJVAZIZ/Y@public.gmane.org>
2016-04-26 21:08 ` Sagi Grimberg
[not found] ` <571FD8D3.1060403-NQWnxTmZq1alnMjI0IkVqw@public.gmane.org>
2016-04-26 23:10 ` Bart Van Assche
2016-04-26 21:06 ` Sagi Grimberg
2016-05-03 9:33 ` Christoph Hellwig
[not found] ` <20160503093307.GL19931-jcswGhMUV9g@public.gmane.org>
2016-05-03 21:13 ` Bart Van Assche
[not found] ` <5729147C.5000904-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-03 21:23 ` Laurence Oberman
2016-05-04 9:09 ` Christoph Hellwig
[not found] ` <20160504090908.GA14787-jcswGhMUV9g@public.gmane.org>
2016-05-04 16:17 ` Bart Van Assche
2016-05-04 23:44 ` Bart Van Assche
[not found] ` <572A8975.9060606-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-05 14:06 ` Christoph Hellwig
[not found] ` <20160505140616.GA15000-jcswGhMUV9g@public.gmane.org>
2016-05-05 19:50 ` Bart Van Assche
2016-05-05 19:58 ` Laurence Oberman
[not found] ` <2102017158.34060891.1462478304470.JavaMail.zimbra-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2016-05-05 20:14 ` Bart Van Assche
[not found] ` <572BA98F.4090109-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-05 20:17 ` Bart Van Assche
[not found] ` <572BAA6F.1060607-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-05 20:55 ` Laurence Oberman
[not found] ` <804753248.34153510.1462481756460.JavaMail.zimbra-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2016-05-05 21:40 ` Bart Van Assche
[not found] ` <572BBDC0.5070205-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-05 22:01 ` Laurence Oberman
[not found] ` <1133135399.34245490.1462485666601.JavaMail.zimbra-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2016-05-05 22:46 ` Bart Van Assche
[not found] ` <572BCD59.7010701-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2016-05-09 22:50 ` Laurence Oberman
[not found] ` <169904890.34764265.1462834206901.JavaMail.zimbra-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2016-05-09 22:55 ` Bart Van Assche
2016-05-05 20:44 ` [PATCH 0/11] More SRP patches for kernel v4.7 Doug Ledford
[not found] ` <f406cd39-c6dc-d458-0ab0-c400fdd5b243-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2016-05-05 20:50 ` Bart Van Assche
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=571A94AF.7000609@sandisk.com \
--to=bart.vanassche-xdaiopvojttbdgjk7y7tuq@public.gmane.org \
--cc=dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
--cc=hch-jcswGhMUV9g@public.gmane.org \
--cc=linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
--cc=loberman-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
--cc=sagi-NQWnxTmZq1alnMjI0IkVqw@public.gmane.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).